基于双字耦合度支持向量机模型的中文文本分类技术研究

Research on Chinese Text Classification Based on Coupling Degree of Double Character Support Vector Machine Model

提出基于双字耦合支持向量机方法对电力客服文本进行分类,由于电力客户投诉口语中包含歧义词较多,所以首先对歧义词进行权重计算,再通过支持向量机对结果进行分类模型识别,效果显示分类效果明显高于普通支持向量机。

This paper proposes a double word coupling support for text power support vector machine method based on market share, because the power customer complaints in spoken English contains ambiguous words more, so the first to calculate the weight of ambiguous words, then the support vector machine to classify the results of model identification, results show that classification effect was significantly higher than that of conventional support vector machine.